level ground, (4) easy accessibility, and (5) no evidence of recent distur- 

 bance, A 10- by 10-meter grid at each site was measured and marked off by 

 corner stakes. 



The terrestrial sweep net sampling method (Table 3) was adapted from 

 Davis and Gray (1966) . The net was vigorously swept back and forth across 

 the upper parts of the vegetation through an horizontal arc of about 1 

 meter. Following each sweep, one step was taken and the direction of the 

 net was reversed. Four samples, each consisting of 20 strokes (10 in each 

 direction), were obtained, one along each edge of the perimeter of the 

 grid. 



After each sample, the contents of the net were placed in a large 

 ethyl acetate-charged killing jar and later transferred to a wide-mouth 

 specimen jar. The samples were cooled in an ice chest for processing in 

 the laboratory where they were then stored in a cold room until the damp 

 and sometimes succulent plant debris could be removed. The insects were 

 sorted and stored dry except for soft-bodied species which were preserved 

 in 70-percent isopropanol solution. 



At each marsh grid, four randomly preselected points were sampled by 

 the clip-quadrat method (Table 3) . The vegetation was first clipped off 

 15 centimeters above the ground. The remaining vegetation was then sliced 

 off at the ground level with a sharp knife and placed in a heavy plastic 

 bag along with any plant litter that could be gathered at the base of the 

 plant. Roots were not collected. Insects seen crawling on the ground 

 inside the quadrat frame were also deposited in the bag. The bags were 

 inflated and securely fastened to avoid crushing the collected plants and 

 insects. The inflated bags were packed in an ice chest for transport to 

 the laboratory. In the laboratory, the plant material was processed in a 

 Berlese-Tullgren apparatus for 7 days. The insects were preserved in 

 small specimen jars filled with 70-percent isopropanol solution. 



b. Debris Line . Invertebrate life of a 40- by 1 -meter (approximate) 

 debris line in the low sand marsh was sampled using the small enclosure 

 method (Table 3) . Four randomly chosen areas in the line were sampled by 

 pushing the small enclosure through the debris (principally eelgrass) and 

 removing the enclosed plants and invertebrates. The samples were processed 

 in the same manner as the clip-quadrat samples. 



All the terrestrial samples were sorted in a flat container under a 

 binocular dissecting scope. Terrestrial sweep net samples, which often 

 contained considerable plant debris, were sorted in a white enamel pan. 

 Samples processed in the Berlese-Tullgren apparatus were sorted in a petri 

 dish. Larvae and the animals less than 0.5 millimeter were not included 

 in the data. 



c. Pan. Several samples were taken in pans in immature and mature 

 high marsh using the aquatic sweep net method (Table 3) . Some laboratory 

 observations of living animals were also made. 



d. Tidal Creeks . Tidal creeks were sampled using small corer, large 

 corer, and aquatic sweep net methods along transects in the mature high 

 marsh in Netarts Bay (1 November 1978) and sedge marsh in Siletz Bay 



20 



